0

Prove that if $f(x)=x$ for rational $x$, and $f(x)=-x$ for irrational $x$, then $\lim\limits_{x \to a} f(x)$ does not exist if $a \neq 0$.

The solution to this problem is intuitively clear, and the solution from the solution manual provides one such descriptive solution. I'd like to know what a more rigorous solution would look like.

My complete solution considers four cases for $a \neq 0$, though I only present one below. The other three are analogous. There is also the case where $a=0$.

$$f(x)=\begin{cases} x, \text{ } x \in \mathbb{Q} \\ -x, \text{ } x \in \mathbb{Q'} \end{cases}$$

$$x \in \mathbb{Q} \implies |f(x)-l|=|x-l|$$ $$x \in \mathbb{Q'} \implies |f(x)-l|=|-x-l|=|-(x+l)|=|x+l|$$

Case 1: $a>0$, $l>0$

If $x \in \mathbb{R}:a-\delta<x<a+\delta$, $x \in \mathbb{Q'}$ then $|f(x)-l|=|x+l|$

Let $x_+ \in \mathbb{Q'}:0<a<x_+<a+\delta$.

Then $|f(x_+)-l|=|x_++l|>|a+l| \geq |a|$

Therefore, we have shown that

$$a>0,l>0, \forall \epsilon>0: 0<\epsilon<a, |x-a|<\delta \implies\exists x:|f(x)-l|>a>\epsilon$$

$\implies \lim\limits_{x\to a}f(x)=l$ is false $\forall l$.

Case 2: $a>0$, $l<0$ This proof is analogous to Case 1's proof, but we look at a $x_+ \in \mathbb{Q}$ instead of $\mathbb{Q'}$.

Case 3: $a<0$, $l>0$ Analogous proof, but we look at an $x_- \in \mathbb{Q}:a-\delta<x_-<a$

Case 4: $a<0$, $l<0$ Analogous proof, but we look at an $x_- \in \mathbb{Q'}: a-\delta<x_-<a$

Case 5: $a=0$, $l=0$

Let $|x|<\epsilon$, $\epsilon>0$.

$$x \in \mathbb{Q} \implies |f(x)-l|=|x|<\epsilon$$ $$x \in \mathbb{Q'} \implies |f(x)-l|=|x|<\epsilon$$

Therefore

$$\forall \epsilon>0, |x|<\epsilon \implies |f(x)|<\epsilon$$

$$\implies \lim\limits_{x \to 0} f(x)=0$$

xoux
  • 4,913
  • Does Spivak use the notation $\mathbb{Q}' = \mathbb{R}\setminus\mathbb{Q}$? I've never seen this before. – Snaw Dec 22 '21 at 12:17
  • I think the simplest way would be to use the Heine definition of a limit, i.e. using sequences. Did the book cover this yet? If so, just take a sequence converging to $a\ne 0$ along the rationals and another one along the irrationals. – Snaw Dec 22 '21 at 12:19
  • No, that notation is used in my solution only. Spivak writes "irrational", as in the original question at the beginning of the post. – xoux Dec 22 '21 at 12:19
  • This Chapter 5 covers limits. Previous chapters cover general properties of numbers, functions, properties of graphs of functions, polar coordinates, conic sections, and vectors. So no sequences. – xoux Dec 22 '21 at 12:20
  • 2
    I found a discussion about a notation for irrational numbers: https://math.stackexchange.com/q/450524/334795 – Hermis14 Dec 22 '21 at 12:20
  • I started using $\mathbb{Q'}$ after searching online for what to use in my proofs. I guess its not a common notation? – xoux Dec 22 '21 at 12:22
  • 2
    It's best using $\mathbb{R}\setminus\mathbb{Q}$, that's the most common one. Anything else you would need to clarify what the notation means. I haven't yet went through your cases 1-4, but why is case 5 needed? The question only asks for $a\ne 0$. – Snaw Dec 22 '21 at 12:23
  • True, it asks us to prove that the limit does not exists if $a \neq 0$. No need to prove that it does exist for $a=0$. – xoux Dec 22 '21 at 13:56

0 Answers0